Abstract In cloud manufacturing, manufacturing business processes are virtualized and encapsulated into manufacturing services. Enterprises can customize their manufacturing business applications on cloud manufacturing platform. To facilitate the customization, a global manufacturing business process is generally built and transformed into a feature model. However, since most global manufacturing business processes have a large number of nodes and complex dependency relationships, it is difficult to transform them into feature models. Therefore, we propose a novel complex manufacturing business process decomposition approach in cloud manufacturing. The approach is beneficial to construct a desired feature model. Firstly, combined with the domain engineering theory, the complex manufacturing business process customization approach in cloud manufacturing is elaborated. Secondly, the complex manufacturing business process decomposition approach is proposed. A global manufacturing business process network (GMBPN) is constructed according to the input and output relationships of manufacturing business activities. Based on the GMBPN, a two-phase decomposition algorithm is presented. In the community detection phase, the nodes in the GMBPN with functional similarity are clustered into the same community with an adaptive strategy. In the sub-process reconstruction phase, the corresponding sub-process is reconstructed in each community. We evaluate the decomposition algorithm with an actual elevator design business process. Experimental results show the proposed algorithm is effective to find the optimal decomposition results.

中文翻译：

---

一种新型的云制造复杂制造业务流程分解方法

摘要 在云制造中，制造业务流程被虚拟化并封装成制造服务。企业可以在云制造平台上定制自己的制造业务应用。为了方便定制，通常会构建一个全球制造业务流程并将其转化为特征模型。但是，由于大多数全球制造业务流程具有大量节点和复杂的依赖关系，因此很难将其转化为特征模型。因此，我们提出了一种新的云制造复杂制造业务流程分解方法。该方法有利于构建所需的特征模型。首先，结合领域工程理论，阐述了云制造中的复杂制造业务流程定制方法。其次，提出了复杂制造业务流程分解方法。根据制造业务活动的输入输出关系构建全球制造业务流程网络（GMBPN）。提出了一种基于GMBPN的两阶段分解算法。在社区检测阶段，将 GMBPN 中功能相似的节点通过自适应策略聚类到同一个社区中。在子流程重建阶段，在每个社区重建相应的子流程。我们通过实际的电梯设计业务流程评估分解算法。